I don't consider having more dialogue options to be the same. I see them as completely perpendicular to the argument I'm making, as dialogue is just another tool for solving problems.
At its core, dialogue is a tool for navigating a chain of evaluated expressions
(if-else blocks)
Whether you use this chain to simulate dialogue or skill checks is a secondary thing.
Continuing from that, in essence we are arguing over input methods. You'd like RPGs to be more like the text-based adenture games, instead of a CYOA hyperlinked structure.
I dunno if you're only discussing preference, you're also discussing feasability and priorities seems to me. And I feel like often the Codex will discuss two things in isolation debating the merits of either without ever looking at the thing relative to the genre itself (a good example is difficulty, where 'dexers will often say something is "isn't hard" but if you follow their line of thinking it turns out they think all RPGs aren't hard so their statement about that specific game is useless in terms of finding out if it's easy or hard relative to other games).
Let's take rusty's healing-someone-in-the-world example. To make a game that generally "works like that" you have to utilize a dedicated and expansive development effort to make sure all the assets of your systems that *aren't* dialogue (or at least, enough for it to feel like a way to interact with the game) can be used in a logical manner. You have to:
1) Design abilities, stats and skills around this
2) Always have it in mind/check for it when you look at "stuff in your world" (e.g. someone who needs healing)
Additionally, all instances of this require unique coding and development effort, or at least the time it requires to figure out "ok how do I do *insert interaction* without additional coding. Use a fire spell to cut down trees, use wind to blow the sails of a boat, use a jump spell to jump through a window, use gust of wind to delight a torch; all these things require different approaches to how the thing is actually implemented in the game.
Meanwhile, solving it through dialogue is a simple matter of CHECK FOR STAT/SKILL/ABILITY->ENABLE DIALOGUE STRING. So you can completely streamline the entire development processes around this. Obviously this means more interactions with less development effort.
Now, the question isn't necessarily: is one of these systems better than the other? It is also: is it
good enough to warrant the additional development effort AND if so, what should be less prioritized than today?
I thought Darklands and to an extend PoE and Deadfire got it right with the CYOA-stuff. Expand how these kinds of systems can enable interaction and role-playing instead of spending your time on sim-stuff. Sim-stuff will always ruin your day because it's a ton of effort for very little gain that will be experienced by exponentially less players than the alternatives (because you have to "guess" a developer's intent or what interactions they allow for, since inevitably a lot of logical interactions won't be in the game).
Lastly, I feel like sim-systems always end up feeling even MORE constraining than abstract dialogue-stuff, for exactly the same reason GTA and Skyrim aren't very immersive to me. When you go for 360 degree verisimilitude you draw more attention to the things that don't work like they should (in our discussion, interactions that seem obvious to you but that the developers didn't bother to implement). Whereas in games that realize they are pixels and not real life you don't expect the game to suddenly explain why something doesn't work like it should, you already waived that part of the fiction contract.
unless you think Deus Ex is worse than a CYOA book I guess.
I don't think you're going to get games with the necessary budgets to make Deus Ex + turn-based open world RPG.
It sounds like a great game. But a game with a budget that it wouldn't be able to deliver on considering the audience for such a thing.
